High Methylmercury in Arctic and Subarctic Ponds ... - ACS Publications

Jun 1, 2015 - In the High Arctic, small ponds have been identified as important ... Climate warming and rising permafrost temperatures are increasing ...
0 downloads 0 Views 3MB Size
Article pubs.acs.org/est

High Methylmercury in Arctic and Subarctic Ponds is Related to Nutrient Levels in the Warming Eastern Canadian Arctic Gwyneth A. MacMillan,† Catherine Girard,† John Chételat,‡ Isabelle Laurion,§ and Marc Amyot*,† †

Centre d’études nordiques, Département de sciences biologiques, Université de Montréal, Montreal, Quebec Canada, H2V 2S9 Environment Canada, National Wildlife Research Centre, Ottawa, Ontario Canada, K1A 0H3 § Centre d’études nordiques, Institut national de la recherche scientifique, Centre Eau, Terre et Environnement, Québec, Quebec Canada, G1K 9A9 ‡

S Supporting Information *

ABSTRACT: Permafrost thaw ponds are ubiquitous in the eastern Canadian Arctic, yet little information exists on their potential as sources of methylmercury (MeHg) to freshwaters. They are microbially active and conducive to methylation of inorganic mercury, and are also affected by Arctic warming. This multiyear study investigated thaw ponds in a discontinuous permafrost region in the Subarctic taiga (Kuujjuarapik-Whapmagoostui, QC) and a continuous permafrost region in the Arctic tundra (Bylot Island, NU). MeHg concentrations in thaw ponds were well above levels measured in most freshwater ecosystems in the Canadian Arctic (>0.1 ng L−1). On Bylot, ice-wedge trough ponds showed significantly higher MeHg (0.3−2.2 ng L−1) than polygonal ponds (0.1−0.3 ng L−1) or lakes ( c a > c, b > c n.s. a>c a>b>c a>b>c

F

p - value

post-hoc

(11.4−16.3) (5.7−7.8) (26−514) (8.34−12.73) na (6.8−18.3) na na (0.74−12.87) (53.6−519.7) (1.51−47.82) (5.2−65.3) (208−804) (0.46−4.76) (3.47−11.16) (0.11−1.24) (1.3−12.9)

50.95 n.s. 8.77 33.73 n.s. n.s. 39.78 28.05 6.03 n.s. n.s. 30.53 n.s. 50.30 16.23 9.26 18.33

b n.s. n.s. b>a>c n.s. b>a>c c>b>a b > a; b > c b > c, b > a

Kuujjuarapik temp (°C) pH cond (μS cm‑1) DO (mgL‑1) A320 (m‑1) DOC (mgL‑1) CO2 (μM) CH4 (μM) SO42‑ (mgL‑1) Fe (μgL‑1) Mn (μgL‑1) TP (μgL‑1) TN (μgL‑1) Chla (μgL‑1) THg (ngL‑1) MeHg (ngL‑1) MeHg (%)

17.8 6.6 52 7.76 31.1 8.71 61.0 0.44 0.39 357.9 6.86 53.7 409 5.91 2.12 0.14 6.7

(14.4−24.2) (5.8−7.2) (27−204) (2.49−9.81) (12.9−53.7) (4.0−28.0) (33.9−141.6) (0.24−1.41) (0.05−12.52) (45.9−2462.3) (1.15−32.40) (15.3−237.3) (228−2899) (1.97−14.30) (0.75−4.35) (0.02−3.56) (2.7−81.9)

7.6 6.2 188 0.41 48.3 7.3 376.6 42.12 0.37 141.4 8.09 184.1 360 52.50 3.66 0.99 27.2

(4.9−13.1) (5.9−6.9) (145−265) (0.41−9.31) (19.7−106.9) (4.2−11.9) (106.9−815.5) (0.48−311.9) (0.08−12.47) (31.6−512.3) (0.61−30.44) (48.1−431.8) (267−496) (7.4−203.4) (1.38−8.56) (0.13−3.07) (6.4−78.1)

13.5 6.9 77 10.41 na 12.5 na na 2.31 186.4 5.52 14.69 530 1.39 6.50 0.33 5.1

a

Surface water geometric mean values (bold) and ranges (min−max) are shown. One-way ANOVA results and post-hoc pairwise comparisons between the 3 groups (Tukey’s HSD) are given (α < 0.05). P-values were corrected for multiple tests (Holm correction) and non-significant tests are shown by n.s. Several variables were not available for the Bylot lakes and for taiga/rock ponds (na). Variables include water temperature (Temp), pH, conductivity (Cond), dissolved oxygen (DO), absorption coefficient of dissolved organic matter at 320 nm (a320), concentrations of dissolved organic carbon (DOC), carbon dioxide (CO2), methane (CH4), major ions, total phosphorus (TP), total nitrogen (TN), chlorophyll a (Chla), total mercury (THg), methylmercury (MeHg), and the percentage of MeHg to THg.

concentrations were estimated by the difference between THg and MeHg concentrations at each site. Normalized data were used to perform all analyses with the R statistical package (R Development Core Team; http://cran.r-project.org). For comparisons of limnological properties and mercury concentrations, the geometric mean (GM) was calculated to better measure the central tendency, calculated as the antilog of the mean of the logarithmic values of the data set. Sites with missing data for multiple variables were not included in the regression analysis, and replacement values were calculated for four sites (each with one missing variable) by imputing the overall variable mean for the type of sample site. Comparisons of limnological properties and mean mercury concentrations were conducted with one-way ANOVAs followed by posthoc pairwise comparisons using the Tukey HSD correction (α < 0.05). Sensitivity analysis with a nonparametric approach (Kruskal−Wallis chi-squared, χ2 rank sum tests) was conducted to test the assumptions of the

for TORT-2 lobster hepatopancreas, National Research Council of Canada) and the mean (±SD) recovery was 99.5 ± 8.4% (n = 97). All water samples were run in duplicate with a Relative Standard Deviation (RSD) of